hadoop-common-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Marc Harris <mhar...@jumptap.com>
Subject Re: Restrictions on the string that can be used for the hadoop key
Date Tue, 22 Jan 2008 17:28:54 GMT
Do you know if there is a bug in the hbase bug tracking system about
this?
a) I don't see this restriction specified in any documentation.
b) It's quite bad that a client error should take the server down
c) It's really bad that it seems to corrupt the data.

- Marc


On Tue, 2008-01-22 at 09:23 -0800, Mike Forrest wrote:

> I ran into a similar error, and it ended up being due to the colon (":") 
> character in the URL.   Encoding the URL or just stripping the colon 
> character(s) fixed it.
> 
> Marc Harris wrote:
> > Are there any restrictions on the string that can be used as the key for
> > an hbase row? I ask because I am using strings of the form:
> >     <URL><space><alphanumerics><comma><numerics>
> >
> > and I frequently get problems that seem to start with the following log
> > message (in my regionserver log file):
> >
> > 2008-01-18 17:24:21,512 FATAL org.apache.hadoop.hbase.HRegionServer: Set
> > stop flag in regionserver/0:0:0:0:0:0:0:0:60020.splitOrCompactChecker
> > java.lang.IllegalArgumentException: java.net.URISyntaxException: Illegal
> > character in scheme name at index 7:
> > hregion_pagefetch,http://someurl.com/fo/bar%20abcd1,5538225121025076292
> >
> > This region server then appears to shut down, and restarting everything
> > (hbase and all hadoop processes) still fails with that same error. I end
> > up having to re-format the entire hadoop directory.
> >
> > Can anyone shell some light on what may be happening? It looks to me
> > like something is adding the prefix "hregion_" to the beginning of my
> > key, and something else is interpreting the whole thing as a URL and
> > getting very confused.
> >
> > Thanks.
> >
> >   
> 

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message